RAMAN SUPPLEMENT


help decide if the product is genuine or not. This is the method that is usually employed with handheld spectrometers. Although this provides good results and is easily implemented, chemometric tools are now more in demand for counterfeit detection since they provide a more accurate answer. Indeed the models that are developed are based on multiple reference


“A correlation calculated between the spectra of suspect and


reference products can help decide if the product is genuine or not”


samples and calibrated so that slight differences between the suspect product and the references can be detected. Several methods have been tested and published, such as Principal Component Analysis (PCA) combined with Hierarchical Cluster Analysis (HCA)16


.


A model based on Support Vector Machines (SVM)10 is presented as an example of quick


identification. In this method, summarised in Figure 3 (page 6), a non linear classification technique, the SVM, was computed on 31 types


FIGURE 2 Raman spectra of a suspect capsule, the genuine reference and the API. The spectra of the suspect and genuine capsules do not present the same profile and API peaks are not observed in the suspect product. This capsule is a counterfeit


or ‘product families’ of genuine capsules and tablets. A suspect product is thus classified by SVM among these 31 product families. For some families, the formulation of the suspect product,


i.e. its closest API dosage to the genuine product, is also provided by SVM or correlation. Once the best match of product family, and possibly formulation, has (have) been identified,


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22